Wednesday, January 17, 2007

Iraqi Dead

With another couple car bombs killing a few hundred more Iraqis today and yesterday, I'm all the more interested in a visit next month to Carleton by one of the statisticians who worked on the study in the Lancet which estimated that about 650,000 Iraqis had died as a result of the war.

I didn't know much about the study beyond the facts that its estimated death rate was far higher than other estimates and that the numbers were instantly politicized by the Left and the Right. Helped out the college library, I did a bit of reading on the methodology and conclusions, and I came away impressed.

First and most important, the methodology yielded not a single figure of Iraqi dead, but a range and a likelihood:

The Lancet study, designed by researchers at Johns Hopkins University in Baltimore, Maryland, is based on a survey conducted between May and July by a team of 10 Iraqi health workers...The team visited 47 neighborhoods in 18 different regions across the country, going door-to-door and asking families about recent deaths. They collected data from a total of 1849 households containing 12,801 residents. For the 14 months before the invasion, the Iraqi families reported 82 deaths, an annual death rate of 5.5 per 1000 people. Within the same households, 547 people died between the start of the invasion and July of this year--an annual increase of 7.8 deaths per 1000. By applying this rate to the entire population of 27 million, the researchers conclude that 655,000 more Iraqis have died than would have if the invasion had never happened. About 8% of these extra deaths are attributed to deteriorating public health, but an estimated 601,000 are violent--56% from gunshots and about 13% each from air strikes, car bombs, and other explosions. The researchers calculate a 95% probability that the true number of violent deaths lies between 426,369 and 793,663.
The probability rating (which is very high) and the range (which is both very wide and extremely high) are crucial pieces of information that don't usually get cited. Short story: it is far more likely than not that about half a million Iraqis have died due to the war.

There is, as you'd expect, some reasonable dissent to this set of conclusions, disagreement, which the Science article describes. There was also some unreasonable, flippant dissent, like this from our Dear Leader:
I don't consider it a credible report... I do know that a lot of innocent people have died, and that troubles me and it grieves me... This report is one -- they put it out before. It was pretty well -- the methodology is pretty well discredited.
Ah, but not so, Mr. President. Guess which giant, rich, and influential institution uses the selfsame "cluster surveying method"? Let Les Roberts, one of the study's co-authors, tell you:
This cluster survey approach is the standard way of measuring mortality in very poor countries where the government isn’t very functional or in times of war. And when UNICEF goes out and measures mortality in any developing country, this is what they do. When the U.S. government went at the end of the war in Kosovo or went at the end of the war in Afghanistan and the U.S. government measured the death rate, this is how they did it. And most ironically, the U.S. government has been spending millions of dollars per year, through something called the Smart Initiative, to train NGOs and UN workers to do cluster surveys to measure mortality in times of wars and disasters.
In other words, it's a good enough technique when we want to figure out how many people the Bad Guys killed, but it's "pretty well discredited" when it's used to figure out how many people the U.S. has killed - or been indirectly responsible for killing.